Content uploaded by Lionel Peter Robert
Author content
All content in this area was uploaded by Lionel Peter Robert on Nov 17, 2024
Content may be subject to copyright.
Chapter 5
Artificial Intelligence, Employee
Engagement, Fairness, and Job Outcomes
Introduction
The use of artificial intelligence (AI) to manage employees is becoming increas-
ingly popular. In this chapter, AI is defined as the ability of a computer system
to sense, reason, and respond to the environment. Computer systems with
advanced AI can engage in sensing, reasoning, and responding in the most com-
plex and dynamic environments. Other terms used to describe the use of AI to
manage employees include algorithm management (O’Connor, 2016). These sys-
tems are being adapted by organizations at a rapid pace to help manage their
workforce (Gerlsbeck, 2018; Kolbjørnsrud, Amico, & Thomas, 2016). The
reason for their popularity is twofold. First, organizations now have access to
huge amounts of data (i.e., big data) about their business operations which can
be leveraged to make more efficient and effective management decisions.
Second, advances in AI now afford organizations the ability to capture and pro-
cess this data in real-time. Organizations can now incorporate the latest informa-
tion into their decision making even in the most complex and dynamic
competitive markets. Despite this, management through AI also presents new
challenges to employees who are now both directed and held accountable by AI.
Employee engagement is essential to the health and productivity of an organi-
zation, yet it is a great challenge faced by organizations globally. Recent research
by Gallup Management Journal reveals that 29% of employees are actively
engaged, 54% are not engaged, and 17% are disengaged (Rao, 2017). In light of
these issues compounded by the introduction and integration of new technologies
such as AI into the workplace, focusing on best practices and tools that enable
employees to bring a full range of cognitive, emotional, and physical energies
into their work roles is of critical importance (Shuck, Adelson, & Reio, 2017).
This chapter seeks to identify the challenges faced by organizations that
employ these new AI systems as they relate to employee engagement and offer
suggestions to meet those challenges. This chapter defines employee engagement
as the degree to which employees are motivated and passionate about their
work. To accomplish this, the chapter provides several things: it presents the
basic underlining concepts behind employee engagement, the factors needed to
ensure that AI promotes employee engagement are discussed trust and risk,
fairness, the important technology characteristics, and issues of AI control and
Employee Engagement
What Is Employee Engagement?
The many definitions of employee engagement available sometimes results in
conflated and confused meanings (Saks, 2006; Shuck, Osam, Zigarmi, & Nimon,
2017). All seem to derive from the core definition of engagement. According to
Merriam Webster, engagement is defined as “tending to draw favorable atten-
tion or interest.”Other definitions include “to occupy the attention or efforts of
a person or persons”or “to attract and hold fast”(see Dictionary.com). Similar
definitions have been applied to employee engagement. Rothbard (2001) defined
it as a psychological presence of attention and absorption. One of the earliest
definitions offered by Kahn (1990) defines it as when an employee applies him/
herself physically, cognitively, and emotionally toward their work. More
recently, HRD scholars have defined the terms as a “positive, active, work-
related psychological state operationalized by the maintenance, intensity, and
direction of cognitive, emotional, and behavioral energy”(Shuck et al., 2017,
p. 269). Taken together, all definitions involve some form of motivation and
affect toward that motivation. In other words, employees are not only willing to
allocate effort and time toward their work but do so in part because they like
doing so. Hence, this chapter defines employee engagement in terms of both
motivation and passion.
Motivation
Motivation is the “internal forces that direct, energize, and sustain work-related
effort”(Chen & Gogus, 2008, p. 290; Kanfer 1990). Motivation represents the
degree to which someone is willing to exert and sustain effort (Chen, Kanfer,
DeShon, Mathieu, & Kozlowski, 2009). Humans are motivated to exert effort
toward achieving a goal or objective (Kanfer, 1990). Motivational theories help
to depict how the roles of goals and objectives can elicit human effort through-
out an activity (Kanfer & Ackerman, 1989; Parker, Bindl, & Strauss, 2010).
Employee motivation can be viewed as the willingness of an employee to
exert effort toward accomplishing his or her work. Employee motivation is a pri-
mary driver of both individual and work group performance (Chen & Kanfer,
2006). In fact, the level of motivation and corresponding effort an employee
exerts have strong impacts on performance (De Jong & Elfring, 2010). This has
been particularly highlighted in the literature investigating social loafing or with-
holding effort (Alnuaimi et al., 2010). Motivation is also positively associated
with job satisfaction and organizational and work group retention. (Chen,
Sharma, Edinger, Shapiro, & Farh, 2011; Hu & Liden, 2014; Park et al., 2010).
In all, motivation is key determinant of important employee outcomes.
62 Managing Technology and Middle- and Low-skilled Employees
retention.
its impact on employee outcomes such as job satisfaction, meaningfulness, and
2010). Goal striving reflects the amount of effort an employee allocates toward
achieving that objective or set of objectives (Parker et al., 2010). Self-efficacy or
the belief that one can achieve a given objective influences the amount of effort
and persistence individuals allocate toward achieving their objective (Parker
et al., 2010). In sum, self-efficacy, goal choice, and goal striving make up the
motivational process that explains the effort an individual allocates toward
achieving a goal (Klein et al., 2008). AI can support each of these processes.
Artificial Intelligence Employee Controls
To fully understand employee engagement and why it is important one should
be familiar with theory X and theory Y of management. Theory X and theory Y
of management represent two fundamental approaches to employee engagement
(Carson, 2005). According to McGregor, theory X seeks to promote engagement
by directing, monitoring, and rewarding and/or punishing the actions of employ-
ees, while theory Y seeks to promote engagement by promoting worker satisfac-
tion through greater job autonomy and freedom and less supervision (Carson,
2005; Kopelman, Prottas, & Davis, 2008). Both are valid approaches to worker
engagement and have pros and cons. For better or for worse, AI management
systems are based on theory X. AI management systems seek to promote worker
engagement by directing, monitoring, and rewarding and/or punishing employ-
ee’s actions. This approach is especially true for sharing economy platforms like
Uber that enable the gig economy (Marquis et al., 2018). But, they can also be
true for AI management systems in traditional organizations.
Organizational controls are used to get workers to act in a way that furthers
the goals and objectives of the organization (Cardinal, Kreutzer, & Miller,
2017). Behavior and outcome controls are two of the most widely used organiza-
tional controls (Cardinal et al., 2017). Behavioral control involves direct, per-
sonal surveillance of a worker’s activities (Dennis et al., 2012). Outcome control
involves an objective measurement or evaluation of the output of a worker’s
activities (Ouchi & Maguire, 1975). Many platform companies employ AI man-
agement systems to control their workers. For example, the Uber app driven by
AI tells the driver who to pick up, how to get there, and where to drop them off
(Rosenblat & Stark, 2016). Given the importance of organizational control in
organizations, it is to be expected that control through AI management systems
is likely to be very important.
Artificial Intelligence Behavior Control
AI behavior control entails the electronic directing and monitoring of an
employee’s work activities to make sure that they are complying with a pre-
defined work standard. Generally, behavior controls are effective when both the
organization and the employees know exactly how a task should be completed.
Artificial Intelligence and Job Outcomes 63
and self-belief in goal attainment (Chen & Gogus, 2008). Goal choice is the pro-
cess of selecting which objective or set of objectives to pursue (Parker et al.,
Motivation is comprised of three core components: goal choice, goal striving,
tive attitudinal outcomes (Cardinal et al., 2017; Jaworski, Stathakopoulos, &
Krishnan, 1993; Sihag & Rijsdijk, 2018). AI behavior control is often enacted
through digital systems.
Artificial Intelligence Outcome Control
AI outcome control is enacted by measuring objective measurements of the
employee’s performance. Behavior control requires directly observing workers’
activities, but outcome control involves evaluating workers’performance after
the service is completed. Examples of outcome controls include monthly or
quarterly sales targets and performance evaluations (Eisenhardt, 1985; Ouchi,
1979). Uber’s rating system is an example of AI outcome control enacted
through a digital platform. Uber drivers who fail to receive a performance rating
of below 4.6/5 can be suspended from driving for Uber.
Artificial Intelligence Trust and Perceived Risk
Artificial Intelligence Trust
Trust in an organization can be defined as an employee’s willingness to be vulner-
able to the actions of the organization (Dirks & Ferrin, 2001; McAllister, 1995).
Based on this definition we define trust in an AI management system as an
employee’s willingness to be vulnerable to the actions of the AI management sys-
tem. Employees should not only rely on their AI management system but do so
with a positive expectation about the outcome. Both the willingness to be vulnera-
ble and the assumption of a beneficial outcome explain why trust will be a crucial
component in the effectiveness of any AI management system. Social exchange
theory is often used to explain the benefits of trust (Dirks & Ferrin, 2001).
According to social exchange theory, employees will take part in an activity
only if they believe the result of their activity is likely to be satisfactory (Blau,
1964; Homans, 1961). For example, Uber drivers are more likely to follow the
guidance provided by the AI management system when they believe that doing
so will lead to a satisfactory outcome. Trust becomes more important when the
level of uncertainty regarding the outcome increases (Robert et al., 2009). This is
because when the outcome is uncertain, individuals do not know whether their
actions will be rewarded with a positive outcome. For example, employees are
less likely to follow the recommendations of the AI management system if there
is a risk that doing so will not lead to a satisfactory outcome. Trust counteracts
such risk by increasing the perceived certainty regarding the likelihood of a satis-
factory outcome (Zand, 1972). The more employees trust the AI management
system, the more willing they will be to take the advice from the system. Trust
reduces anxiety and fears by allowing employees to rule out undesirable yet pos-
sible outcomes. For example, when Uber drivers are asked to relocate to pick up
64 Managing Technology and Middle- and Low-skilled Employees
interact with store customers to ensure they are following company policies.
Behavior control has been shown to lead to high intrinsic motivation and posi-
For example, a retail store manager may monitor the way employees greet and
Artificial Intelligence Perceived Risk
Perceived risk can be defined as the degree of uncertainty associated with an out-
come (Sitkin & Pablo, 1992). Perceived risk differentiates trust from the beha-
viors that require trust (Mayer, Davis, & Schoorman, 1995). It is what
determines whether an individual will translate trust beliefs into trusting actions.
Organizations often minimize perceived risk by enacting organizational control
systems like sanctions for violating trust or institutional structures to minimize
losses (Sheppard & Sherman, 1998; Sitkin & Pablo, 1992). Individuals tend to be
risk-averse (Friedman & Savage, 1948) and weigh the risk of loss more heavily
than the potential benefits (Tversky & Kahneman, 1974). Different people might
also have different perceptions of risks when presented with the same situation.
Employees are likely to associate perceived risk with marginal gains or losses
associated with a given AI management system. For example, Uber drivers have
some expectations about the amount of money they can make on a given trans-
action. They likely compare this to the amount of money they believe they are
likely to make when following the guidance provided by AI management
system. For example, if a driver has been asked to relocate to another location
to serve a future rider, the driver might ask herself, “Will I make more money
relocating than the money I am sure to make here?”or “Why am I being asked
to relocate?”If drivers believe that the monetary outcome provided by AI man-
agement system is below their expected outcome, they are likely to ignore the
system, either in the short or the long term.
Artificial Intelligence Fairness
The AI management system must convince employees that it is fair and justice.
Generally, fairness can be viewed as the framework to explain employees’trust
toward their organization (Saunders & Thornhill, 2003). An employee’s level of
commitment to an employer is directly related to their belief that their employer
treats them fairly (Hendrix, Robbins, Miller, & Summers, 1998). Perceived fair-
ness is derived from the organizational justice literature, which is based on
Adams’(1965) equity theory. Equity theory states that individuals believe that
the distribution of rewards should be based on an individual’s contribution.
In organizations, equity typically refers to the expected inputs by an employee
relative to the outcomes the employee expects to receive from his or her employer
(Colquitt, 2012). Employees compare their input and outcomes to that of
other employees and expect to see roughly the same input-to-outcome ratio
(Cowherd & Levine, 1992; Hendrix et al., 1998). When a discrepancy is detected,
employees will attempt to resolve it in one of three ways: (1) altering their percep-
tions of the inputs and outcomes, (2) changing their inputs, or (3) leaving the orga-
nization (Cowherd & Levine 1992; Hopkins & Weathington, 2006). There are
three widely used types of perceived fairness: procedural, distributive, and
Artificial Intelligence and Job Outcomes 65
Uber’s AI management system will likely lead to a positive outcome.
possible riders, the drivers are much more likely to do so when they believe
Artificial Intelligence Procedural Fairness
Procedural fairness is the perceived fairness of the processes used by the organi-
zation (Hendrix et al., 1998: Holbrook, 1999; Saunders & Thornhill, 2003).
These processes manifest themselves in how fair employees believe that the AI
management system decision process is to each employee. Procedural fairness
will be low (1) when employees believe the AI management system decision pro-
cesses are inconsistent or biased and/or (2) whether employees can express their
concerns regarding such decisions (Leventhal, 1980). For example, procedural
fairness for employees managed by AI systems would represent the way com-
plaints are handled, how their performance is evaluated, or their ability to voice
their concern over new policies.
Artificial Intelligence Distributive Fairness
Distributive fairness refers to the perceived fairness of outcomes and allocation
of resources (Greenberg & Colquitt, 2013). These outcomes and resources are
normally in the form of pay or praise. When an employee believes the AI man-
agement system leads to employees receiving equal pay for equal work, distribu-
tive fairness should be high. However, when employees believe their use of the
AI management system either leads to or exacerbates pay inequalities, distribu-
tive fairness should be low.
Artificial Intelligence Interactional Fairness
Interactional fairness relates to the treatment employees receive as decisions are
made, and this can be broken into interpersonal and informational justice
(Colquitt, Conlon, Wesson, Porter, & Ng, 2001). Interpersonal fairness can be
defined as the degree of respect employees are provided, while informational
fairness refers to the information provided regarding why procedures are carried
out and why outcomes are distributed (Colquitt, 2012). For employees being
managed by AI management systems, interaction fairness would refer to what
degree they think the AI respects them and provides an explanation for its
decisions.
Artificial Intelligence Technology Characteristics
Perceived ease of use and usefulness are two important technology characteris-
tics. They have been found to be important predictors of technology use
(Venkatesh, Morris, Davis, & Davis, 2003). More recently, they have been used
to predict the use of digital platforms involving electronic commerce sites such
as eBay (Gefen, Karahanna, & Straub, 2003; Pavlou, 2003). Their importance
has also been tied to trust in the customers’use of AI management systems
66 Managing Technology and Middle- and Low-skilled Employees
determine if AI management systems are based on each of these three types.
interactional fairness (Greenberg & Colquitt, 2013). Employees are likely to
Artificial Intelligence Perceived Ease of Use
Perceived ease of use is defined as the degree employees believe the system is
straightforward and relatively trouble-free to use (Venkatesh et al., 2003).
Systems high in ease of use require little effort to employ and are not very diffi-
cult to learn. Employees will use systems more often when they believe they will
experience little difficulty in using them (Venkatesh & Davis, 2000). When sys-
tems are not easy to use, anticipated difficulties often act as psychological bar-
riers to employing systems (Venkatesh, Thong, & Xu, 2012). As such, employees
managed by AI systems would be less likely to engage with AI management sys-
tems if they found it difficult to interact with. On the other hand, employees
managed by AI systems would be more willing to engage them if they believed it
to be easy to use.
Artificial Intelligence Perceived Usefulness
Perceived usefulness is defined as the degree employees believe that using the sys-
tem will benefit them (Venkatesh et al., 2012). This normally pertains to either
better job performance or a decrease in the time needed to perform the same
amount of work. When employees believe a system is useful, they are much
more likely to employ the system to perform their work (Venkatesh et al., 2003).
In the case of AI, we should expect drivers to trust and employ AI when they
believe it is useful. We should expect the opposite when they believe the system
is not useful.
Artificial Intelligence, Job Satisfaction, Job Meaningfulness,
and Retention
Throughout the literature on the management of employees three outcomes are
often considered vital to understanding if management practices are successful:
job satisfaction, job meaningfulness, and employee retention. Job satisfaction
can be defined as the degree of positive attitude workers have toward their job
(Beer, 1964). Job meaningfulness is the degree to which someone feels that his or
her job is worthwhile, useful, and valuable (Kahn, 1990). Employee retention is
a measure of how many employees are retained by an organization (Carsten &
Spector, 1987; Iverson & Pullman, 2000). Taken together, all three represent
important measures of employee outcomes.
Research has shown that relationships and interactions with others in the
workplace are vital to promoting job satisfaction, meaningfulness, and retention
(Karasek, Triantis, & Chaudhry, 1982; Tett & Meyer, 1993). These relationships
and interactions include informal conversations with colleagues and formal feed-
back and mentoring from supervisors (Bateman, 2009; Fay & Kline, 2011). In
Artificial Intelligence and Job Outcomes 67
to better understand trust and its implications for the use of AI management
systems.
(Xiao & Benbasat, 2007). Therefore, we include both ease of use and usefulness
with other humans (Taneja et al., 2011). Yet, the use of AI systems could possi-
bly reduce or eliminate such social interactions.
How can organizations with employees managed primarily by AI systems
promote social interactions? One approach is to build online communities on
digital platforms. This approach is similar to many current communities spon-
sored by companies like Uber or Airbnb. Organizations can set up their own
online communities to allow employees to interact not only with other employ-
ees but also with the company. These online communities can both build “com-
munity”among employees and allow companies to contact and assess the needs
of their employees. Of course, issues of privacy should not be overlooked.
Organizations must seek and receive approval from employees before monitor-
ing their online actions. That withstanding, this approach can help replace tradi-
tional social interactions which seem lost in these new digital AI-enabled work
arrangements.
Conclusion
Artificial Intelligence is used to manage employees through employee engage-
ment. This occurs by motivating employees and controlling their actions.
However, issues of trust, perceived risk, and fairness play a vital role in deter-
mining whether such systems will be effective at managing employees over the
long run. In addition, AI-driven systems must be easy to use and be viewed as
useful if companies hope to encourage their use. This chapter presents and dis-
cusses these issues along employee outcomes such as job satisfaction, meaning-
fulness, and employee retention.
68 Managing Technology and Middle- and Low-skilled Employees
supervisors is often the primary driver of employee outcomes like job satisfac-
tion, meaningfulness, and retention (Bateman, 2009; Carsten & Spector, 1987;
Ross, 2005). This is because humans are social beings who need interactions
fact, social support derived from such interactions with both colleagues and